This concluding chapter is divided into two sections. In the opening section, I summarize my results and address some of the implications and limitations of my analysis. In Section 7.2, I describe my agenda for future research.

SUMMARY OF RESULTS, IMPLICATIONS, AND LIMITATIONS OF ANALYSIS

In this book, I have attempted to advance the debate over capital flow liberalization in the developing world by studying the determinants of lax bank regulation under liberal capital flows. Aside from cronyism and gridlock, which have previously been identified, I identified a third causal path to lax regulation, namely, the path of incredible signaling. I showed that if a chief executive does not have the freedom to appoint an official who shares his preferences as a signaler of confidential bank regulatory information, the outcome will be an incredible long-term commitment to stringent bank regulation. Using the tools of game theory, I showed that the difference in preferences between a signaler and the chief executive that results in serious miscommunication is very small in the realm of bank regulation. I showed that miscommunication results in an incredible commitment to stringent regulation even in the presence of a chief executive who does not have close ties to bankers, by causing the chief executive to miscalculate his responses to shocks to the banking sector.

I then identified ways to solve signaling problems, such as appointing friends or close associates who share the chief executive's regulatory priorities to senior financial positions.

In this chapter, I show how a common understanding has emerged that lax bank regulation presents immense dangers for countries operating under liberal capital flows. I begin by presenting the neo-classical case for capital flow liberalization. I then describe the attack on this case following the Asian crisis. Finally, I describe how even prominent proponents of liberalization now accept that the success of liberalization may be contingent on stringent bank regulation.

Prior to the 1980s, most developing countries maintained a significant body of regulations limiting the inflow and outflow of capital across national borders. Foreign exchange transactions had to be approved by government officials and were subject to stringent limits, domestic banks were tightly restrained from borrowing from private sources overseas, and stock markets faced significant legal obstacles to accessing international funds. Starting in the 1980s, the International Monetary Fund (IMF) began to place immense pressure on developing countries to dismantle these and other barriers to the inflow and outflow of international capital. Four arguments, which are sometimes jointly referred to as the neo-classical case for liberalization, were offered in justification for this pressure.

First, it was argued that environments with liberal capital flows, referred to in short as environments with open capital accounts, would help developing countries gain access to funds from developed countries. This would enable them to achieve investment levels that exceeded their domestic savings rates, and thus help them grow faster in the long run.

In this chapter, I address the three countries in the sample that were democracies. At some point in time each of these countries had a chief executive without crony links to bankers, who was forced to operate with a signaler/signalers with regulatory preferences that were different from his own thanks to the checks on his power. On each occasion there is evidence of serious problems in communication that went unresolved, resulting in lax regulation. This, of course, is in line with the predictions of the theory.

THAILAND

Thailand proved to be an extremely lax regulator of banks in the wake of capital flow liberalization in the early 1990s. Prior to liberalization, however, Thailand's central bank was not considered to be an extremely lax regulator. To place Thailand's post-liberalization regulatory performance in context, I begin with a background section that addresses the political determinants of bank regulation in Thailand from the 1960s to liberalization. In Section 4.2, I describe the political environment in which Chuan Leekpai, the Thai prime minister for the three years following capital flow liberalization, operated. In Section 4.3, I describe Chuan's inner circle of banking advisors. In Section 4.4, I describe Thailand's weak record of enforcing bank regulations during Chuan's term, and Section 4.5 addresses the regulatory performance of Chuan's successors.

Background: Bank Regulatory Governance in Thailand Before 1992

In the three decades prior to the liberalization of capital flows in 1992–93, Thailand was predominantly governed by a succession of authoritarian or semi-authoritarian regimes, that is, regimes with very few checks on the chief executive's power.

In this chapter, I demonstrate why a small difference in preferences between the chief executive and financial bureaucrats renders signals vague. I also show why the threshold difference in preferences between the chief executive and signalers that generates completely uninformative signals is relatively low in the realm of bank regulation. I begin by defining some terms. I then describe the logic that links a significant difference in preferences between the chief executive and the central bank governor to lax regulation, when the governor is the sole signaler of information collected by bank supervisors. In Section 3.3, I address the scenario in which both the central bank governor and the finance minister are signalers.

DEFINING SOME TERMS

Banking Sector Robustness

The signaling models presented here focus on a key aspect of a banking sector's robustness: the degree to which a banking sector is capitalized relative to expected loan defaults. I focus on this aspect of robustness for the signaling model because signaling considerations are important in realms where some actors have private information, and this condition is often likely to be met when it comes to the relationship between capitalization and expected defaults, for the reasons listed in Chapter 1.

As per the principles of bank accounting, a bank that has a shareholder capital buffer that is insufficiently large to cover loan defaults is considered to be insolvent.

The prediction of this book is that authoritarian environments are unlikely to be subject to either signaling or gridlock problems. Thus, unlike environments with even a moderate number of checks, we should generally observe outcomes that are consistent with the preferences of the chief executive. In this chapter I address the two countries where the chief executives adopted unorthodox solutions to the signaling problem following capital flow liberalization. In one case, the chief executive appointed a long-time friend from his home village to closely monitor the central bank and finance ministry from a special senior advisory position. In the other, the chief executive sidelined senior financial bureaucrats and instead relied on relatives and close cronies who owned banks as his primary source of information on the banking sector. In neither case is there evidence of signaling problems. The chief executives of both countries had close ties to the banking sector. The outcome, lax regulation, was in line with both chief executives' preferences.

MALAYSIA

In the years leading up to the Asian crisis, Malaysia was an authoritarian country where political power was largely concentrated in its prime minister, Mahathir Mohamad. In Section 5.1.1, I provide some background information on Malaysia's political and economic environment in the years leading up to the liberalization of capital inflows in 1990. In Section 5.1.2, I describe Prime Minister Mahathir's inner circle of banking advisors in the 1990s. In Section 5.1.3, I address the bank regulatory environment during this period.

Assume, strictly for the time being, that the chief executive has a posterior belief that the signaled value of z is true only when the messages sent by both signalers agree. Assume, also strictly for the time being, that when the messages disagree he believes that z lies between ϖ − 2xs1 and ϖ + 2xs1. (At the end of this section, I demonstrate that these beliefs are consistent with the senders' strategies.)

Given the above posterior beliefs, the chief executive will find it optimal to choose k = z when the messages agree. This is the case because, given that x = k − z, k = z will yield x = 0, which is his ideal point. Both senders will prefer to signal the true value of z, if the chief executive's choice when the messages do not agree will yield a value of x further from both their ideal points than xc. Because ω is uniformly distributed, and given that when the messages disagree the chief executive believes that it lies between ϖ − 2xs1 and ϖ + 2xs1, the chief executive's choice of k when the messages disagree is ϖ. (This choice maximizes his expected utility when z lies between these values, because it minimizes the expected distance of x from his ideal point.) Given that x = k − z, the chief executive's choice of k = ϖ when the messages disagree yields outcomes that are to the left of −2xs1 when z < ϖ + 2xs1 and to the right of 2xs1 when z < ϖ − 2xs1.

We now live in a world where capital often moves freely across national borders. In this world, developing countries have increasingly been subjected to devastating financial crises caused by the sudden withdrawal of foreign capital. These crises have had severe humanitarian consequences; in East Asia in 1997–98 some countries experienced economic contractions comparable to levels seen in the Great Depression. How do such crises come about? I focus on a novel causal path, that of miscommunication. I demonstrate why developing democracies are exceptionally vulnerable to breakdowns in communication between financial officials and the chief executive. These breakdowns have disastrous consequences because they result in inadequate bank regulation, which encourages the withdrawal of foreign capital.

This book contributes to three literatures. The first is the literature on globalization of capital. This literature has hitherto paid little attention to how globalization can have disastrous consequences in political environments where there are problems in the credible communication of financial information, and this is a contribution of my book. The second is the literature on the politics of financial crises. It is plausible that the presence of ill-informed chief executives raises the likelihood of financial crises. However, scholars of crises have hitherto been unable to systematically identify where we are likely to observe an ill-informed chief executive, thanks to the absence of any preexisting analytical framework for such an analysis.

In this chapter, I address the two authoritarian countries where chief executives adopted the orthodox solution to the signaling problem, namely, to appoint close associates who shared their preferences to key bureaucratic positions. In both cases, the regulatory outcomes were in line with the chief executives' preferences for stringent regulation.

SINGAPORE

Recall that an authoritarian country whose chief executive has arm's length relations with the banking sector is predicted to have a stringent bank regulatory environment, unimpeded by signaling problems or gridlock. I show in this chapter that, although Singapore changed its development strategy several times in the course of the last three decades, there was one constant: arm's length relations between the chief executive and the entire business community, including the banking sector. I begin with a background section that addresses the period between 1965 and 1990, when Lee Kwan Yew served as the chief executive. In Section 6.2, I describe the political environment under Lee's successor, Goh Chok Tong. In Section 6.3, I describe the bank regulatory environment in Singapore in the years leading up to the Asian crisis.

Background: Singapore Between 1965 and 1990

Ever since its emergence as an independent state, Singapore has operated under a system of government that falls well short of being a democracy. The government's powers include detention without trial, deregistration and replacement of radical unions with compliant ones, and withdrawal of licenses from newspapers deemed to be opposed to national interests.

In 1997, prior to the implementation of post-crisis reforms, the World Bank evaluated the bank regulatory environments of twelve countries on six criteria known as CAMELOT: capitalization, loan classification, management, liquidity, operating environment, and transparency. The countries were rank ordered in each of these areas, and their rank scores were added up to yield an overall score. Countries were then assigned to five different categories of stringency based on the clustering of points, with the top two categories being indicative of stringent regulation. Of the twelve countries, seven were the Asian countries with liberal capital flows that are considered in this book, while the remaining five were Latin American countries. Table A.1 shows how these countries ranked on each of the components of regulation. Note that of the two democracies that had stringent bank regulatory environments, the commitment of one (Argentina) collapsed in the late 1990s. As for the second, Chile, the country's regulatory record was achieved under a regime with substantial capital controls, meaning that we have no way of knowing if it would have been robust to massive unrestricted capital inflows. (Chile's score on the IMF's capital controls index was 0.89 out of a possible 1 in 1996, with 1 indicating a closed environment. By way of comparison, the most closed economy of the ones studied in this book was South Korea, with a score of 0.7.)

A detailed description of each of the regulatory measures is available in Caprio (1998).

A liberal democracy rules by the mandate and wishes of its citizens. Periodic elections place candidates into legislative and executive offices based on some voting rule, such as a simple majority or plurality of votes cast. The political system may be two-party majoritarian, proportional representation, or some similar system. The “liberal” adjective underscores that the system preserves the civil and political rights of citizens and foreign residents (see Doyle, 1997). In a well-functioning liberal democracy, the political and civil rights of the minority are protected by the ruling government. The press is allowed to report the news, and everyone has the right to express his or her views. As a result, people can criticize the government and its policies without fear of reprisals. Suspected criminals have civil rights – for example, to be charged with a crime, to obtain counsel, and to have a fair trial. Election results are tallied in an open and accurate fashion. When a government loses an election, it relinquishes office and allows for a peaceful and orderly transition of power.

An essential requirement of a liberal democracy is the protection of its people's lives and property. A government that fails to provide this security will lose support and be voted out of office. If, for example, a liberal democracy is unable to control a terrorist campaign that murders innocent individuals on city streets or on public transit, then the government will appear inept and lose popularity.

In a globalized world with a high volume of cross-border flows, transnational terrorism is a global public bad, while action to control or eliminate it is the quintessential global public good. As such, antiterrorism efforts yield nonrival benefits – enhanced security – received by all at-risk countries. The formation of far-flung terrorist networks has greatly increased the spatial dispersion of benefits derived from measures against these networks. The theory of public good supply teaches that as the dispersion of these measures' benefits increases, their underprovision worsens as providers fail to include the benefits that their efforts confer on others when deciding upon antiterrorism actions (see Chapter 4; Sandler, 1997, 2004).

The sheer volume of cross-border exchanges of all types makes it possible to monitor but a small fraction, thereby affording opportunities for terrorists to move personnel and equipment internationally. For example, over a half-billion people cross US borders each year. Also, 130 million motor vehicles, 2.5 million railcars, and 5.7 million cargo containers transit US borders annually (White House, 2004, p. 165). The combination of globalization and technological advances means that even the most secure borders may be penetrated by determined terrorists who utilize technologies (for example, communication advances) and apply innovative methods to circumvent security upgrades. Terrorists weigh relative risks to identify the least secure venue or weakest link at which to stage their attack against a targeted nation's assets.